Personal Blogs

p453... Patterns are descriptive findings. Lots of respondents found this

Themes are more abstract than this, there is an element of interpretation. Lots of respondents found this... and I will label that group as this abstract theme.

Inductive analysis is about discovering patterns

Deductive analysis is when the data are analysed according to an existing framework.

Can i be inductive within the boundaries of a framework? Recognising that the questions and slant of interviews were influenced by a framework of prior knowledge and reading, but within that, attempting to let the story in the data emerge for themselves?

Grounded theory emphasises being immersed in the data - grounded - so that embedded meanings and relationships can emerge.

p454

Once patterns, themes and/or categories have been established thorugh inductive analysis, you can move to deductive analysis  in testing and affirming the authenticity and appropriateness of the inductive content analysis. So, you can move from one to the other. I guess it is a form of triangulation. Looking at what you think you see from other perspectives

An "interplay of making inductions (deriving concepts, their properties and dimensions from data) and deductions (hypothesizing about the relationships between concepts) (Strauss and Corbin 1998, in Patton 2002)

HOWEVER

Analytic induction begins with an analyst's deduced propositions or theory-derived hypotheses and is a procedure for verifying theories and propositions based on qualitative data".

Nope, that's not what I'm doing. I'm using those propositions and hypotheses as a tool for exploring qualitative data, the conceptual model is a way of making a boundary so that there is an acceptably small amount of data to handle and already organised in some way, but not then verifying. Verifying to me means truth-checking and I am not doing that.

Later in the same paragraph however, it does seem more to describe what i want to do:

Sometimes [...] qualitative analysis is first deductive or quasi-deductive and then inductive as when, for example, the analyst begins by examining the data in terms of theory- derived sensitizing concepts or applying a theoretical framework developed by someone else [...] After or alongside this deductive phase of analysis, the researcher strives to look at the data afresh for undiscovered patterns and emergent undertandings (inductive analysis).

Inductive analysis is one of the primary characteristics of qualitative inquiry, so we need strategies for thinking and working inductively.  Here are two:

1. identify, define and elucidate the categories developed by they people studied (emic)

2. seeing patterns that he people studied do not describe in their own terms, so the analyst develops them (etic)

OK... steps

1. get organized. do an inventory. is everything complete? holes in data? is everything labeled? Dates, places, ways of identifying sources

2. read through the lot

3. read through the lot again

4. Now what about Computer Assisted Qualitative Data Management and Analysis? Helpful or a distraction? Will the time invested in understanding it be more than offset by its power in helping understand patterns in the data? As a further output of this course familiarity with a CAQDMA could be useful. p444 Patton gives some examples of software programs

5. I can glimpse already patterns and systems coming forth out of the data i have so far. I think certain ways of seeing the purpose will be linked to certain other ways of seeing the alumni program. Is that something a CAQDMA can help me see? Is there any difference by country and batch and academic level? (Obviously not statistical but question-raising differences?)

p 445 The 4 fundamental types of information that contribute to the construction of a finding of "answer" in qualitative analysis are:

1. characteristics of the sources where information is sought
2. primary information or objects collected from the sources
3. secondary information or objects created to aid in th interpretation of primary objects, and
4. characteristics of the coders who construct the secondary objects (MacQueen and Milstein 1993:31, in Patton)

I think what this is saying is that you have information about the source, then the info itself, your interpretation of that, and your framework that you are operating in.

So then you segment and create metadata. Segments are bits of text, metadata are categories, codes, comments, annotations, graphical representations

I am feeling this tension. I really want to be Mrs Inductive - it is an approach which fits how I see the world and how I would like to treat other people. For me it seems more respectful almost to treat people's unique thoughts and data as fresh and deserving of an analysis that brings forth something new.

And yet...

My questions for the interviews and in the questionnaires were based on a pretty thorough lit review. And my pattern-loving mind already got in there before I spoke to even one fellow and came up with a nice elegant conceptual model. so the questions are based around that. so i guess the responses will - surprise surprise! - fit the model.

Learning? Advancement? Zero.

So now... I am thinking about "inductive analysis" (Patton 2002) and I am thinking about other approaches i heard mention of once on a video "template analysis" and "framework analysis". Maybe there is something there to help me break open some new learning.

And that is partly the fault of my personality type - I love Closure (it's the J in my ENFJ) - so I have to fight that. Patton says that the qualitative analyst's main tool is the analyst herself. So i have to make sure that tool is as self-aware as possible by finding ways to prevent premature closure. I will have to do things like reading all the transcripts while sitting on my hands so i have to read the lot before i am allowed to take notes of themes!